This paper presents a unified approach to Chinese statistical language modeling (SLM). Applying SLM techniques like trigram language models to Chinese is challenging because (1) t...
m, modules, types and operations), different kinds of abstractions (functional/data, types/objects etc.) without falling into a loose collection of diagram languages. Considering a...
The supremacy of n-gram models in statistical language modelling has recently been challenged by parametric models that use distributed representations to counteract the difficult...
Recently, the bag-of-words approach has been successfully applied to automatic image annotation, object recognition, etc. The method needs to first quantize an image using the vis...
Incremental natural language understanding is the task of assigning semantic representations to successively larger prefixes of utterances. We compare two types of statistical mod...